AI use map
What is AI-only, AI-assist, or human-first in the workflow you send.
INSTRUMENT 003 · SELF-CHECK
AI is already in your workflow. What should stay yours?
Shiori helps you notice where AI is useful, where it may be quietly taking over judgment, and which boundary to try before speed turns into drift.
OWNERSHIP SIGNAL
Keep these human-first.
- ReviewCan this be checked fast?
- VoiceDoes it still sound like you?
- JudgmentWho owns the final choice?
- OwnershipIs review more than approval?
- MeaningShould this move slower?
SYMPTOM DOORS
Start from the sentence that already feels true.
Pick one door to seed the self-check, or answer from scratch.
SHIORI SELF-CHECK
Find your current AI drift layer.
Local only. No login, email, network call, or saved result.
SILENT DIAGNOSIS
Send your AI workflow. Get a written boundary diagnosis.
No workshop required. No teaching required. Shiori starts as a low-pressure written review of where AI is helping, where it may be replacing judgment, and what should remain human-first.
Contact through Bright RavenJudgment-risk zones
Where review, voice, ownership, or meaning can become a rubber stamp.
Boundary map
AI-only / AI-assist / human-first boundaries written in plain workflow language.
3 workflow changes
Three concrete changes to test before adding another tool or automation layer.
RESEARCH SNAPSHOT
Shiori is not competing on speed or audit readiness.
Automation sells speed
Workflow consultants usually ask what can be automated, ranked by time saved or ROI.
Governance sells audit readiness
Formal AI governance asks whether controls, evidence, and oversight are ready for review.
Shiori focuses on what stays yours
Judgment, voice, review, meaning, and ownership are treated as design boundaries.
WHY THIS EXISTS
The reading map is the deeper theory, not the front door.
Below is Shiori's research base: predictive processing, judgment quality, and human-AI decision making. Use it when you want the theory behind why evidence, friction, voice, and ownership belong in the same workflow conversation.
INTERACTIVE MAP
三軸 constellation
目前顯示:星圖。點任一節點查看 detail。
CROSS-AXIS LINKS
跨軸連線
桌面星圖會畫線;手機與鍵盤使用者可直接操作下方 chips。
DESIGN CONSEQUENCE
AI assist,not AI-first.
Beacon
給設計師 example-based evidence:跟過去某個被罰案例對照,比只丟 WCAG ratio 更能養成 internalized intuition。
shikigarasu
Verdict + Findings 容易被 rubber-stamp。若改成 past-case comparison,更符合「給 evidence、不給結論」。
Noise 的反論
noise reduction 是 AI 主導的強論點。這份 reading map 不迴避它,而是把它放進設計哲學的壓力測試。